CN110998463B

CN110998463B - Knowledge recommendation server and method for defect inspection

Info

Publication number: CN110998463B
Application number: CN201880007391.6A
Authority: CN
Inventors: 方伟; 郑楚发; 简儒浩; 王怡颖; 陈诗丛; 廖建民; 李川; 郭朝晖; 黄邦瑄; 赖绍伟; 徐世宗
Original assignee: ASML Holding NV
Current assignee: ASML Holding NV
Priority date: 2017-01-18
Filing date: 2018-01-15
Publication date: 2023-08-25
Anticipated expiration: 2038-01-15
Also published as: KR20220019063A; TWI709182B; US20190362488A1; TWI793455B; KR102468184B1; US11650576B2; KR102357310B1; TW202117883A; TW201841276A; WO2018134158A1; KR20190103407A; CN110998463A

Abstract

A knowledge recommendation server for defect inspection. The server includes a processor electronically coupled to an electronic storage device that stores a plurality of knowledge files related to wafer defects. The processor is configured to execute the set of instructions to cause the server to: receiving a request for knowledge recommendation for inspection of the inspection image from the defect classification server; searching a knowledge file matched with the inspection image in the electronic storage device; and transmitting the search result to the defect classification server.

Description

Knowledge recommendation server and method for defect inspection

Cross Reference to Related Applications

The present application is based on and claims priority from U.S. provisional application No.62/447,565, filed on 18 of 2017 and entitled "Auto Loading Knowledge for Defect Review" and U.S. provisional application No.62/612,593, filed on 31 of 12 of 2017 and entitled "Knowledge Recommendation for Defect Review", and the disclosures of both applications are incorporated herein by reference in their entirety.

Technical Field

The present disclosure relates generally to systems and methods for recommending knowledge files for defect inspection.

Background

With the continued increase in demand for low-cost and high-performance electronic devices, it is critical to control the manufacturing process of integrated circuits included in the electronic devices so as to reduce the number of defects that affect yield. The defect inspection process has been integrated into the control of the manufacturing process. During the defect inspection process, defects on a semiconductor wafer are automatically identified and classified into various defect types. Although defect classification may be performed automatically, user intervention is always required to select the knowledge file on which to perform automatic defect classification. User intervention significantly reduces the throughput of the manufacturing process.

Disclosure of Invention

According to some embodiments of the present disclosure, a server for knowledge recommendation for defect inspection is provided. The server includes a processor electronically coupled to an electronic storage device that stores a plurality of knowledge files related to wafer defects. The processor is configured to execute the set of instructions to cause the server to: receiving a request for knowledge recommendation for inspection of the inspection image from the defect classification server; searching a knowledge file matched with the inspection image in the electronic storage device; and transmitting the search result to the defect classification server.

According to some embodiments of the present disclosure, a server for defect classification is provided. The server includes a processor configured to execute a set of instructions to cause the server to: receiving a test image of the wafer from a test tool; sending a request for knowledge recommendation to a knowledge recommendation server; receiving knowledge recommendation results from a knowledge recommendation server; determining whether the knowledge recommendation result includes a knowledge file; and in response to determining that the knowledge recommendation includes a knowledge file, performing automatic defect classification on the inspection image by using the knowledge file.

According to some embodiments of the present disclosure, a defect inspection system is provided. The defect inspection system includes an inspection tool for inspecting a wafer, a defect classification server electronically coupled to the inspection tool, and a knowledge recommendation server electronically coupled to the defect classification server. The defect classification server includes a processor configured to execute a set of instructions to cause the defect classification server to: receiving a test image of the wafer from a test tool; and sending a request for knowledge recommendation for verifying the verification image. The knowledge recommendation server includes a processor configured to execute a set of instructions to cause the knowledge recommendation server to: searching for a knowledge file matching the inspection image in response to receiving a request for knowledge recommendation from the defect recommendation server; and transmitting the knowledge recommendation result to the defect classification server.

According to some embodiments of the present disclosure, a method for knowledge recommendation is provided. The method comprises the following steps: receiving a request for knowledge recommendation for inspection of the inspection image from the defect classification server; searching a knowledge file matched with the inspection image in an electronic storage device, wherein the electronic storage device stores a plurality of knowledge files related to wafer defects; and transmitting the search result to the defect classification server.

According to some embodiments of the present disclosure, a method for defect classification is provided. The method comprises the following steps: receiving a test image of the wafer from a test tool; sending a request for knowledge recommendation to a knowledge recommendation server; receiving knowledge recommendation results from a knowledge recommendation server; determining whether the knowledge recommendation result includes a knowledge file; in response to determining that the knowledge recommendation includes a knowledge file, performing automatic defect classification on the inspection image by using the knowledge file.

According to some embodiments of the present disclosure, a method for defect inspection is provided. The method comprises the following steps: receiving, by the defect classification server, an inspection image of the wafer from the inspection tool; transmitting, by the defect classification server, a request for knowledge recommendation for verification of the verification image to a knowledge recommendation server; searching, by the knowledge recommendation server, a knowledge file that matches the verification image in response to receiving a request for knowledge recommendation from the defect recommendation server; transmitting a knowledge recommendation result to the defect classification server by the knowledge recommendation server; and determining, by the defect classification server, whether the knowledge recommendation result transmitted from the knowledge recommendation server includes a knowledge file; and in response to determining that the knowledge recommendation includes a knowledge file, performing, by the defect classification server, automatic defect classification on the inspection image by using the knowledge file.

According to some embodiments of the present disclosure, a non-transitory computer readable medium is provided. The non-transitory computer-readable medium stores a set of instructions executable by at least one processor of the knowledge recommendation server to cause the knowledge recommendation server to perform a method. The method comprises the following steps: receiving a request for knowledge recommendation for inspection of the inspection image from the defect classification server; searching a knowledge file matched with the inspection image in an electronic storage device, wherein the electronic storage device stores a plurality of knowledge files related to wafer defects; and transmitting the search result to the defect classification server.

According to some embodiments of the present disclosure, a non-transitory computer readable medium is provided. The non-transitory computer-readable medium stores a set of instructions executable by at least one processor of a defect classification server to cause the defect classification server to perform a method. The method comprises the following steps: receiving a test image of the wafer from a test tool; sending a request for knowledge recommendation to a knowledge recommendation server; receiving knowledge recommendation results from a knowledge recommendation server; determining whether the knowledge recommendation result includes a knowledge file; and in response to determining that the knowledge recommendation includes a knowledge file, performing automatic defect classification on the inspection image by using the knowledge file.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate several embodiments.

Fig. 1 is a schematic diagram illustrating an exemplary Electron Beam Inspection (EBI) system consistent with an embodiment of the present disclosure.

Fig. 2 is a schematic diagram illustrating an exemplary electron beam tool, which may be part of the exemplary electron beam inspection of fig. 1, consistent with embodiments of the present disclosure.

FIG. 3 is a block diagram of an exemplary defect inspection system consistent with an embodiment of the present disclosure.

Fig. 4 shows an example of a defective patch (patch) image.

FIG. 5 schematically illustrates an exemplary defect knowledge file consistent with some embodiments of the present disclosure.

FIG. 6 is a flow chart of an exemplary process for knowledge recommendation consistent with an embodiment of the present disclosure.

Fig. 7 is a flowchart of an exemplary process for automatic defect classification consistent with an embodiment of the present disclosure.

FIG. 8 is a block diagram of an exemplary computer system upon which embodiments described herein may be implemented.

Detailed Description

Reference will now be made in detail to the exemplary embodiments illustrated in the accompanying drawings. Although the following embodiments are described in the context of using an electron beam, the present disclosure is not limited thereto. Other types of charged particle beams may be similarly applied.

The disclosed embodiments provide a knowledge recommendation server for use in a defect inspection system. The knowledge recommendation server includes an electronic storage device that stores a plurality of knowledge files related to wafer defects. When a request for knowledge recommendation for verifying the verification image is received from the defect classification server, the knowledge recommendation server searches the electronic storage device for a knowledge file matching the verification image and provides the knowledge file to the defect classification server. Then, the defect classification server performs automatic defect classification on the inspection image by using the knowledge file provided by the knowledge recommendation server.

Fig. 1 illustrates an exemplary Electron Beam Inspection (EBI) system 100 consistent with embodiments of the present disclosure. As shown in fig. 1, the EBI system 100 includes a main chamber 101, a load/lock chamber 102, an electron beam tool 104, and an Equipment Front End Module (EFEM) 106. The electron beam tool 104 is located within the main chamber 101. The EFEM 106 includes a first feed port 106a and a second feed port 106b. The EFEM 106 may include additional feed port(s). The first and second feed ports 106a, 106b receive a wafer cassette containing wafers (e.g., semiconductor wafers or wafers made of other materials) or samples (the wafers and samples are hereinafter collectively referred to as "wafers") to be inspected.

One or more robotic arms (not shown) in the EFEM 106 transfer wafers to the load/lock chamber 102. The load/lock chamber 102 is connected to a load/lock vacuum pump system (not shown) that removes gas molecules in the load/lock chamber 102 to achieve a first pressure below atmospheric pressure. After the first pressure is reached, one or more robotic arms (not shown) transfer the wafer from the load/lock chamber 102 to the main chamber 101. The main chamber 101 is connected to a main chamber vacuum pump system (not shown) that removes gas molecules in the main chamber 101 to reach a second pressure that is lower than the first pressure. After reaching the second pressure, the wafer is subjected to inspection by the e-beam tool 104.

Fig. 2 illustrates exemplary components of an electron beam tool 104 consistent with embodiments of the present disclosure. As shown in fig. 2, the e-beam tool 104 includes a motorized stage 200 and a wafer holder 202 supported by the motorized stage 200 to hold a wafer 203 to be inspected. The electron beam tool 104 further includes an objective lens assembly 204, an electron detector 206 (which includes electron sensor surfaces 206a and 206 b), an objective lens aperture 208, a condenser lens 210, a beam limiting aperture 212, a gun aperture 214, an anode 216, and a cathode 218. In one embodiment, objective lens assembly 204 may include a modified wobble deceleration immersion objective lens (SORIL) that includes a pole piece 204a, a control electrode 204b, a deflector 204c, and an excitation coil 204d. The electron beam tool 104 may additionally include an energy dispersive X-ray spectrometer (EDS) detector (not shown) to characterize the material on the wafer.

By applying a voltage between the anode 216 and the cathode 218, a primary electron beam 220 is emitted from the cathode 218. The primary electron beam 220 passes through the gun aperture 214 and the beam limiting aperture 212, both of which may determine the size of the electron beam entering the condenser lens 210 located below the beam limiting aperture 212. The condenser lens 210 focuses the primary electron beam 220 before the beam enters the objective aperture 208 to set the size of the electron beam before entering the objective assembly 204. The deflector 204c deflects the primary electron beam 220 to support beam scanning of the wafer. For example, during scanning, the deflector 204c may be controlled to sequentially deflect the primary electron beam 220 to different positions of the upper surface of the wafer 203 at different points in time to provide data for image reconstruction of different portions of the wafer 203. In addition, the deflector 204c can also be controlled to deflect the primary electron beam 220 onto different sides of the wafer 203 at different points in time at a particular location to provide data for stereoscopic image reconstruction of the wafer structure at that location. Additionally, in some embodiments, the anode 216 and cathode 218 may be configured to generate a plurality of primary electron beams 220, and the electron beam tool 104 may include a plurality of deflectors 204c to project the plurality of primary electron beams 220 simultaneously to different portions/sides of the wafer to provide data for image reconstruction of different portions of the wafer 203.

The excitation coil 204d and the pole piece 204a generate a magnetic field that begins at one end of the pole piece 204a and ends at the other end of the pole piece 204 a. The portion of the wafer 203 scanned by the primary electron beam 220 may be immersed in a magnetic field and may be charged, which in turn generates an electric field. The electric field reduces the energy of the impinging primary electron beam 220 near the surface of the wafer before the primary electron beam 220 collides with the wafer. A control electrode 204b, electrically isolated from pole piece 204a, controls the electric field across the wafer to prevent micro-arching of the wafer and ensure proper beam focus.

Upon receiving the primary electron beam 220, a secondary electron beam 222 may be emitted from a portion of the wafer 203. The secondary electron beam 222 may form a beam spot (e.g., one of beam spots 240a and 240 b) on the sensor surfaces 206a and 206b of the electron detector 206. The electron detector 206 may generate a signal (e.g., voltage, current, etc.) representative of the intensity of the beam spot and provide the signal to a processing system (not shown in fig. 2). The intensity of the secondary electron beam 222 and the resulting beam spot may vary depending on the external and/or internal structure of the wafer 203. In addition, as discussed above, the primary electron beam 220 may be projected onto different locations on the upper surface of the wafer and/or onto different sides of the wafer at particular locations to generate secondary electron beams 222 (and resulting beam spots) of different intensities. Thus, by mapping the intensity of the beam spot to the location of the wafer 203, the processing system can reconstruct an image that reflects the internal and/or external structure of the wafer 203.

After the wafer image is obtained, the wafer image may be transferred to a computer system, where the system may identify defects on the wafer image and classify the defects into a plurality of categories according to the type of defect. FIG. 3 is a schematic diagram of a defect inspection system 300 consistent with an embodiment of the present disclosure.

Referring to fig. 3, defect inspection system 300 includes a wafer inspection system 310, an Automatic Defect Classification (ADC) server 320, and a knowledge recommendation server 330 electrically coupled to ADC server 320. Wafer inspection system 310 may be Electron Beam Inspection (EBI) system 100 described with respect to fig. 1. It should be appreciated that the ADC server 320 and the knowledge recommendation server 330 may be part of the EBI system 100 and/or remote from the EBI system 100.

Wafer inspection system 310 may be any inspection system capable of generating an inspection image of a wafer. The wafer may be a semiconductor wafer substrate, or a semiconductor wafer substrate having one or more epitaxial layers and/or handle films. Wafer inspection system 310 may be any currently available or under development wafer inspection system. Embodiments of the present disclosure are not limited to a particular type of wafer inspection system 310, as long as it is capable of generating wafer images with a resolution high enough to view critical features (e.g., less than 20 nm) on a wafer.

The ADC server 320 has a communication interface 322 that is electrically coupled to the wafer inspection system 310 to receive wafer images. The ADC server 320 also includes a processor 324, the processor 324 being configured to analyze the wafer image and detect and classify wafer defects appearing on the wafer image by using the defect knowledge file. The defect knowledge file may be provided to the ADC server 320 manually by an operator. Alternatively, the defect knowledge file may be automatically provided to the ADC server 320 by the knowledge recommendation server 330, according to an embodiment of the present disclosure, which will be described in detail below.

Knowledge recommendation server 330 is electrically coupled to ADC server 320. Knowledge recommendation server 330 includes a processor 332 and a storage device 334. The processor 332 is configured to build a plurality of defect knowledge files and store the plurality of defect knowledge files in the storage 334.

The plurality of defect knowledge files contain information related to various types of defects generated during various stages of the wafer fabrication process. The various stages of the wafer fabrication process may include, but are not limited to, photolithography processes, etching processes, chemical Mechanical Polishing (CMP) processes, and interconnect formation processes. Defects generated in the photolithography process may include, but are not limited to, photoresist (RP) residual defects due to PR degradation or impurities, lift-off defects, bridging defects, bubble defects, and dummy pattern deletion defects due to pattern shift. Defects generated in the etching process may include, but are not limited to, etch residual defects, overetch, defects, and open defects. Defects generated in the CMP process may include, but are not limited to, slurry residue defects, dishing defects and erosion defects due to polishing rate variation, and scratches due to polishing. Defects generated in the interconnect formation process may include, but are not limited to, wire breakage defects, void defects, extrusion defects, and bridging defects.

The processor 332 is configured to construct a plurality of defect knowledge files based on the plurality of defect patch images. The plurality of defect patch images may be generated by a wafer inspection tool, such as the electron beam tool 104 illustrated in fig. 2. The defect patch image is a small image (e.g., 34 pixels by 34 pixels) of the portion of the wafer containing the defect. The defect patch image is typically centered on the defect and includes adjacent pixels of the defect.

FIG. 4 illustrates exemplary defect patch images 410-450 of various defects in a metal interconnect layer. Image 410 is a patch image of a broken line defect. Image 420 is a patch image of small void defects, where the dimensions of the voids are smaller than the width of the metal lines. Image 430 is a patch image of a pinch defect. Image 440 is a patch image of a large void defect, where the size of the void is greater than or equal to the width of the metal line. Image 450 is a patch image bridging the defect.

Still referring to fig. 3, the processor 332 may be trained via a machine learning process to construct a knowledge file related to a particular type of defect based on multiple defect patch images of that type of defect. For example, the processor 332 may be trained to construct a knowledge file related to wire break defects generated in the interconnect formation process based on a plurality of defect patch images of the wire break defects.

The processor 332 is further configured to search for a knowledge file that matches the wafer image included in the received request in response to a request for knowledge recommendation from the ADC server 320, and provide the knowledge file to the ADC server 320.

The storage device 334 stores an ADC data center containing a plurality of defect knowledge files related to various types of defects generated during various stages of the wafer fabrication process. A plurality of defect knowledge files in the ADC data center may be constructed by the processor 332 of the knowledge recommendation server 330. Alternatively, a portion of the defect knowledge file in storage 334 may be preset by a user or an external computer system and may be preloaded into storage 334.

The defect knowledge file may include general information about a single type of defect. The general information may include patch images and feature parameters (e.g., size, edge roughness, depth, height, etc.) that will later be used for classification of single type defects. Alternatively, according to some embodiments of the present disclosure, the defect knowledge file may include general information about multiple types of defects present in the same process layer of the wafer. The single process layer may be, for example, a substrate layer, an epitaxial layer, a thin film layer, a photoresist layer, an oxide layer, a metal interconnect layer, and the like.

FIG. 5 schematically illustrates an exemplary defect knowledge file in accordance with some embodiments of the present disclosure. As shown in FIG. 5, knowledge file 510 includes a plurality of sub-knowledge files 520-530 each associated with a single type of defect in a wafer process layer. For example, the wafer process layer is a metal interconnect layer, and knowledge file 510 includes information related to various types of defects in the metal interconnect layer. The sub-knowledge file 520 includes information related to the wire break defect and includes 9 different patch images of the wire break defect and features extracted from the 9 patch images. The sub-knowledge file 522 includes information related to the small void defect and includes 41 different patch images of the small void defect and features extracted from the 9 patch images. The sub-knowledge file 524 includes information related to the pinch-out defect and includes 17 different patch images of the pinch-out defect and features extracted from the 17 patch images. The sub-knowledge file 526 includes information related to the large void defect and includes 8 different patch images of the large void defect and features extracted from the 8 patch images. The sub-knowledge file 528 includes information related to bridging defects and includes 19 different patch images of bridging defects and features extracted from the 19 patch images. The sub-knowledge file 530 includes information related to other types of defects and includes 4 different patch images of other types of defects and features extracted from the 4 patch images.

FIG. 6 is a flow chart of an exemplary process 600 for knowledge recommendation consistent with an embodiment of the present disclosure. The process 600 may be performed by a knowledge recommendation server (e.g., the knowledge recommendation server 330 illustrated in fig. 3). The knowledge recommendation server is wirelessly coupled by a communication cable or through a network to a storage device (e.g., the storage device illustrated in fig. 3) that stores a plurality of knowledge files.

According to some embodiments of the present disclosure, process 600 for knowledge recommendation is performed based on the assumption that defects inspected in the same wafer process layer and under similar inspection conditions may share the same knowledge file. That is, the knowledge file includes information related to various defects that are in the same wafer process layer and inspected under similar inspection conditions. The inspection condition refers to the setting of an inspection tool that generates a plurality of defective patch images. Taking the e-beam tool 104 illustrated in fig. 2 as an example, the inspection conditions may include e-beam size, e-beam energy, scan speed, focus conditions, and the like. Typically, different wafer process layers are inspected under different inspection conditions, while the same wafer process layer is inspected under similar conditions, and the resulting inspection image will have similar features for a single type of defect.

As shown in fig. 6, first, in step 602 to step 608, the knowledge recommendation server pre-processes the knowledge file in the storage device by analyzing the defect patch image and the wafer inspection conditions. Specifically, in step 602, the knowledge recommendation server obtains a plurality of defect patch images from a plurality of knowledge files stored in a storage device. As discussed above, the storage device includes a plurality of knowledge files, and each knowledge file includes information related to a different defect type in the same wafer process layer. The information related to each type of defect includes a defect patch image of the type of defect.

In step 604, the knowledge recommendation server analyzes the defect patch images, extracts feature parameters from the plurality of defect patch images, and normalizes the feature parameters. The feature parameters are parameters describing various features of the defect. For example, the characteristic parameters of the defect may include the size, depth, height, surface roughness, edge roughness of the defect. The knowledge recommendation server may extract the feature parameters directly from the defect patch image. Alternatively, for certain specific types of features, the knowledge recommendation server may calculate corresponding feature parameters based on other parameters extracted from the defect patch image.

At step 606, the knowledge recommendation server builds a plurality of weighted representation models based on the normalized patch image features. Each weighted representation model represents a type of defect and includes a plurality of representative feature parameters that describe the features of the type of defect. Taking the void in the metal interconnect layer as an example of one type of defect, the memory device then stores a plurality of patch images of the void in the metal interconnect layer. Each patch image contains a void. The knowledge recommendation server extracts a set of feature parameters from each patch image. Each feature parameter describes one of a plurality of features (e.g., size, depth, edge roughness) of the void. After extracting feature parameters from all patch images of the void, the knowledge recommendation server calculates, for each feature, a weighted average of feature parameters corresponding to the feature and extracted from all of the plurality of patch images of the void. When calculating the weighted average of the feature parameters, the weights for each feature parameter may be preset by the user. The knowledge recommendation server takes the weighted average of the calculated feature parameters as a representative feature parameter for the feature. For example, the knowledge recommendation server calculates a weighted average of the sizes of the different voids in the different patch images and treats the calculated weighted average as a representative size of the voids in the metal interconnect layer. After computing the weighted average of all features of the gap, the knowledge recommendation server builds a weighted representation model for the gap based on all weighted averages of the feature parameters of the gap. The knowledge recommendation server then repeats the same process to build a weighted representation model for other types of defects.

At step 608, the knowledge recommendation server stores the plurality of weighted representation models in a storage device. For example, the knowledge recommendation server may add the weighted representation model to an existing knowledge file in a storage device. Alternatively, the knowledge recommendation server may construct a plurality of new knowledge files, each of which is composed of a plurality of weighted representation models of the plurality of types of defects in the corresponding wafer process layer. The knowledge recommendation server then saves the new knowledge file in the storage device.

After the preprocessing of steps 602 through 608, the knowledge recommendation server may be used to provide knowledge recommendations. Specifically, at step 610, the knowledge recommendation server receives a request for knowledge recommendation from a defect classification server (e.g., ADC server 320 illustrated in fig. 3). The defect classification server may be coupled to the knowledge recommendation server wirelessly by a communication cable or through a network. The request for knowledge recommendation includes an inspection image of the wafer generated by the inspection tool, and information related to the identified defects in the inspection image. The information related to the defect includes a patch image of the defect and a feature parameter.

In step 612, the knowledge recommendation server accesses the storage unit to search for a knowledge file that matches the defect information in the inspection image. For example, the knowledge recommendation server extracts feature parameters from patch images that verify defects in the images. The knowledge recommendation server then compares the extracted feature parameters with feature parameters stored in a knowledge file in the storage device. The knowledge recommendation server employs a feature combination search strategy to select a knowledge file containing defective feature parameters that best match feature parameters extracted from the inspection image. Alternatively, the knowledge recommendation server may employ a pattern search strategy to search for a knowledge file containing defective patch images that match defective patch images of the inspection image.

At step 614, the knowledge recommendation server sends the search results as knowledge recommendation results to the defect classification server, and process 600 is complete. If the knowledge recommendation server finds a knowledge file that matches defect information identified in the inspection image, the knowledge recommendation server provides the knowledge file to the defect classification server. On the other hand, if the knowledge recommendation server cannot find a knowledge file that matches the defect information identified in the inspection image, the knowledge recommendation server may send a message to inform the defect classification server that the knowledge file cannot be found. Alternatively, the knowledge recommendation server may provide the default knowledge file to the defect classification server.

Alternatively, according to some embodiments of the present disclosure, if the knowledge recommendation server cannot find a knowledge file that matches defect information identified in the inspection image, the knowledge recommendation server generates a new weighted representation model by varying weights for the feature parameters, constructs a new knowledge file using the new weighted representation model, and determines whether the new knowledge file matches defect information identified in the inspection image. The knowledge recommendation server may repeatedly build new knowledge files until the new knowledge files match defect information identified in the inspection image.

Fig. 7 is a flowchart of an exemplary process 700 for automatic defect classification consistent with an embodiment of the present disclosure. Process 700 may be performed by a defect classification server (e.g., ADC server 320 in fig. 3). The defect classification server is wirelessly coupled to the inspection tool (e.g., e-beam tool 104 in fig. 2) and the knowledge recommendation server (e.g., knowledge recommendation server 330 in fig. 3) by a communication cable or through a network.

As shown in fig. 7, first, at step 702, a defect classification server receives a wafer inspection image generated by an inspection system. In step 704, the defect classification server analyzes the inspection image to identify a plurality of defects, extracts information related to each defect, and then sends a request for knowledge recommendation to the knowledge recommendation server. The request for knowledge recommendation includes extracted information related to the identified defect in the inspection image, including a patch image and feature parameters of the defect.

When the knowledge recommendation server receives a request for knowledge recommendation, the knowledge recommendation server searches for a knowledge file that matches defect information included in the request in process 600 and transmits the search result to the defect classification server. Thus, at step 706, the defect classification server receives search results from the knowledge recommendation server.

At step 708, the defect classification server determines whether the search results include a knowledge file that matches defect information for the wafer inspection image. If the search results do not include such knowledge files (step 708: NO), then at step 710 the defect classification server prompts the user to construct a knowledge file based on the wafer inspection image. The defect classification server may prompt the user by displaying a prompt message on the display or by sending a signal to the user terminal. After step 710, process 710 is complete.

If the search result includes a knowledge file (step 708: yes), then the defect classification server performs automatic defect classification by using the knowledge file provided by the knowledge recommendation server at step 712. In automatic defect classification, a defect classification server compares the defect patch image and/or feature parameters for each defect in the wafer inspection image with the defect patch image and/or feature parameters in the knowledge file and identifies the type of defect.

After automatic defect classification, the defect classification server prompts the user to examine the defect classification results at step 714. For example, the defect classification server may display the defect classification result on a display that displays a patch image of the defect in the wafer inspection image with a label indicating the type of defect. In response to the prompt, the user may visually inspect the defect classification result, determine whether the result is accurate, and provide feedback regarding the determination result to the defect classification server. In addition, if the defect classification result is inaccurate, the user may correct the defect classification result and provide feedback regarding the corrected result to the defect classification server. For example, if the defect classification result of one or more defects is inaccurate, the user may input a name of the correct type of the one or more defects.

In step 716, the defect classification server determines whether the defect classification result is accurate. If the defect classification result is accurate (step 716: yes), process 700 is complete. If the defect classification result is inaccurate (step 716: no), then the defect classification server communicates the user's correction result feedback to the knowledge recommendation server at step 718. The knowledge recommendation server may then update the knowledge file in the storage unit with the correction result. After step 718, the process is complete.

FIG. 8 is a block diagram of an exemplary computer system 800 with which embodiments described herein may be implemented. At least one of the knowledge recommendation server and the defect classification server described above may be implemented using the computer system 800.

Computer system 800 includes a bus 802 or other communication mechanism for communicating information, and one or more hardware processors 804 (represented for simplicity as processor 804; e.g., processor 332 of knowledge recommendation server 330 or processor 324 of defect classification server 320 of FIG. 3) coupled with bus 802 for processing information. The hardware processor 804 may be, for example, one or more microprocessors.

Computer system 800 also includes a main memory 806, such as a Random Access Memory (RAM) or other dynamic storage device, coupled to bus 802 for storing information and instructions to be executed by processor 804. Main memory 806 also may be used for storing temporary variables or other intermediate information during execution of instructions to be executed by processor 804. Such instructions, when stored in a non-transitory storage medium accessible to processor 804, cause computer system 800 to be a special purpose machine that is customized to perform the operations specified in the instructions.

Computer system 800 also includes a Read Only Memory (ROM) 808 or other static storage device coupled to bus 802 for storing static information and instructions for processor 804. A storage device 810, such as a magnetic disk, optical disk, or USB thumb drive (flash drive), for example, storage device 334 of knowledge recommendation server 330 of fig. 3, is provided and coupled to bus 802 for storing information and instructions.

Computer system 800 may be coupled via bus 802 to a display 812, such as a Cathode Ray Tube (CRT), liquid Crystal Display (LCD), or touch screen, for displaying information to a computer user. An input device 814, including alphanumeric and other keys, is coupled to bus 802 for communicating information and command selections to processor 804. Another type of user input device is cursor control 816, such as a mouse, a trackball, or cursor direction keys for communicating direction information and command selections to processor 804 and for controlling cursor movement on display 812. The input device typically has two degrees of freedom in two axes, a first axis (e.g., x) and a second axis (e.g., y), which allows the device to specify positions in a plane. In some embodiments, the same directional information and command selections as cursor control may be achieved via receiving a touch on a touch screen without a cursor.

Computing system 800 may include a user interface module to implement a Graphical User Interface (GUI) that may be stored in a mass storage device as executable software code executed by one or more computing devices. By way of example, the module and other modules may include various components, such as software components, object-oriented software components, class components and task components, procedures, functions, fields, programs, subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays, and variables. The modules may include one or more components of the system 300 illustrated in fig. 3, for example.

Computer system 800 may implement the techniques described herein using custom hardwired logic, one or more ASICs or FPGAs, firmware, and/or program logic, in combination with the computer system, to make computer system 800 a or to program computer system 800 into a special purpose machine. According to some embodiments, the operations, functions, and techniques described herein, as well as other features, are performed by computer system 800 in response to processor 804 executing one or more sequences of one or more instructions contained in main memory 806. Such instructions may be read into main memory 806 from another storage medium, such as storage device 810. Execution of the sequences of instructions contained in main memory 806 causes processor 804 to perform the method steps described herein (e.g., process 600 of fig. 6, or process 700 of fig. 7). In alternative embodiments, hard-wired circuitry may be used in place of or in combination with software instructions.

Embodiments may be further described using the following clauses:

1. a server for knowledge recommendation for defect inspection, comprising:

a processor electronically coupled to an electronic storage device storing a plurality of knowledge files related to wafer defects, the processor configured to execute a set of instructions to cause a server to:

receiving a request for knowledge recommendation for inspection of the inspection image from the defect classification server;

searching a knowledge file matched with the inspection image in the electronic storage device; and

the search results are transmitted to a defect classification server.

2. The server of clause 1, wherein the processor is configured to execute the set of instructions to cause the server to:

acquiring a plurality of defect patch images from a knowledge file in an electronic storage device;

extracting characteristic parameters from a plurality of defect patch images and normalizing the characteristic parameters;

generating a plurality of weighted representation models based on the normalized feature parameters, each weighted representation model representing a defect type; and

the weighted representation model is stored in an electronic storage device.

3. The server of clause 1, wherein the request for knowledge recommendation includes feature parameters that verify the plurality of defects identified in the image, and

In searching for a knowledge file that matches the verification image, the processor is configured to execute a set of instructions to cause the server to:

comparing the characteristic parameters of the plurality of defects identified in the inspection image with the characteristic parameters in the knowledge file in the electronic storage device; and

a knowledge file is searched for having characteristic parameters that match characteristic parameters of a plurality of defects identified in the newly obtained inspection image.

4. The server of any of clauses 1-3, wherein each knowledge file includes a plurality of defect patch images and feature parameters of different types of defects in the same wafer process layer and generated by the inspection tool under the same inspection conditions.

5. The server of any one of clauses 1 to 4, wherein the defect patch image is generated by an electron beam inspection tool.

6. A server for defect classification, comprising:

a processor configured to execute a set of instructions to cause a server to:

receiving a test image of the wafer from a test tool;

sending a request for knowledge recommendation to a knowledge recommendation server;

receiving knowledge recommendation results from a knowledge recommendation server;

determining whether the knowledge recommendation result includes a knowledge file; and

In response to determining that the knowledge recommendation includes a knowledge file, performing automatic defect classification on the inspection image by using the knowledge file.

7. The server of clause 6, wherein the processor is configured to execute the set of instructions to cause the server to:

in response to determining that the knowledge recommendation result does not include a knowledge file, the user is prompted to construct a knowledge file based on the verification image.

8. The server of clause 6, wherein the processor is configured to execute the set of instructions to cause the server to:

after performing the automatic defect classification on the inspection image, the user is prompted to check the result of the automatic defect classification.

9. The server of clause 8, wherein the processor is configured to execute the set of instructions to cause the server to:

in response to receiving an input from the user indicating that the result of the automatic defect classification is inaccurate and includes a corrected defect classification result, transmitting the corrected defect classification result to the knowledge recommendation server,

wherein the knowledge recommendation server updates the knowledge file based on the corrected defect classification result.

10. The server of any of clauses 6 to 9, wherein the inspection tool is an electron beam inspection tool.

11. A defect inspection system comprising:

An inspection tool for inspecting the wafer;

a defect classification server electronically coupled to the inspection tool and comprising a processor configured to execute a set of instructions to cause the defect classification server to:

receiving a test image of the wafer from a test tool; and

transmitting a request for knowledge recommendation for verifying the verification image;

a knowledge recommendation server electronically coupled to the defect classification server and comprising a processor configured to execute a set of instructions to cause the knowledge recommendation server to:

searching for a knowledge file matching the inspection image in response to receiving a request for knowledge recommendation from the defect recommendation server; and

and transmitting the knowledge recommendation result to the defect classification server.

12. The defect inspection system of clause 11, wherein the knowledge recommendation server comprises an electronic storage device storing a plurality of knowledge files related to wafer defects.

13. The defect inspection system of clause 12, wherein the processor of the knowledge recommendation server is configured to execute the set of instructions to cause the knowledge recommendation server to:

the weighted representation model is stored in an electronic storage device.

14. The defect inspection system of clause 13, wherein the request for knowledge recommendation includes characteristic parameters of the plurality of defects identified in the newly obtained inspection image, and

15. The defect inspection system of any of clauses 12 to 14, wherein each knowledge file comprises a plurality of defect patch images and feature parameters of different types of defects in the same wafer process layer and generated by the inspection tool under the same inspection conditions.

16. The defect inspection system of any of clauses 11-15, wherein the inspection tool is an electron beam inspection tool.

17. The defect inspection system of any of clauses 11 to 16, wherein the processor of the defect classification server is configured to execute the set of instructions to cause the defect classification server to:

determining whether the knowledge recommendation result transmitted from the knowledge recommendation server includes a knowledge file; and

18. The defect inspection system of any of clauses 11 to 17, wherein the processor of the defect classification server is configured to execute the set of instructions to cause the defect classification server to:

19. The defect inspection system of any of clauses 11 to 17, wherein the processor of the defect classification server is configured to execute the set of instructions to cause the defect classification server to:

20. The defect inspection system of clause 19, wherein the processor of the defect classification server is configured to execute the set of instructions to cause the defect classification server to:

wherein the processor of the knowledge recommendation server is configured to execute the set of instructions to cause the knowledge recommendation server to update the knowledge file in the electronic storage device based on the corrected defect classification result.

21. A method for knowledge recommendation, comprising:

searching a knowledge file matched with the inspection image in an electronic storage device, wherein the electronic storage device stores a plurality of knowledge files related to wafer defects; and

the search results are transmitted to a defect classification server.

22. The method of clause 21, further comprising:

acquiring a plurality of defect patch images from a plurality of knowledge files;

the weighted representation model is stored in an electronic storage device.

23. The method of clause 21, wherein the request for knowledge recommendation includes feature parameters that verify the plurality of defects identified in the image, and

searching the knowledge file further includes:

24. The method of any of clauses 21-23, wherein each knowledge file includes a plurality of defect patch images and feature parameters of different types of defects in the same wafer process layer and generated by the inspection tool under the same inspection conditions.

25. The method of any of clauses 21 to 24, wherein the defect patch image is generated by an electron beam inspection tool.

26. A method for defect classification, comprising:

receiving a test image of the wafer from a test tool;

27. The method of clause 26, further comprising:

28. The method of clause 26, further comprising:

29. The method of clause 28, further comprising:

30. The method of any of clauses 26 to 29, wherein the inspection tool is an electron beam inspection tool.

31. A method for defect inspection, comprising:

receiving, by the defect classification server, an inspection image of the wafer from the inspection tool;

transmitting, by the defect classification server, a request for knowledge recommendation for verifying the verification image to the knowledge recommendation server;

Searching, by the knowledge recommendation server, a knowledge file that matches the verification image in response to receiving a request for knowledge recommendation from the defect recommendation server;

transmitting a knowledge recommendation result to the defect classification server by the knowledge recommendation server;

determining, by the defect classification server, whether the knowledge recommendation result transmitted from the knowledge recommendation server includes a knowledge file; and

in response to determining that the knowledge recommendation includes a knowledge file, performing, by the defect classification server, automatic defect classification on the inspection image by using the knowledge file.

32. A non-transitory computer-readable medium storing a set of instructions executable by at least one processor of a knowledge recommendation server to cause the knowledge recommendation server to perform a method comprising:

the search results are transmitted to a defect classification server.

33. The non-transitory computer readable medium of clause 32, wherein the method further comprises:

the weighted representation model is stored in an electronic storage device.

34. The non-transitory computer-readable medium of clause 32, wherein the request for knowledge recommendation includes feature parameters that verify the plurality of defects identified in the image, and

the method further comprises the steps of:

35. The non-transitory computer readable medium of clauses 32-34, wherein each knowledge file includes a plurality of defect patch images and feature parameters of different types of defects in the same wafer process layer and generated by the inspection tool under the same inspection conditions.

36. The non-transitory computer readable medium of clauses 32-35, wherein the defect patch image is generated by an electron beam inspection tool.

37. A non-transitory computer-readable medium storing a set of instructions executable by at least one processor of a defect classification server to cause the defect classification server to perform a method comprising:

receiving a test image of the wafer from a test tool;

38. The non-transitory computer readable medium of clause 37, wherein the method further comprises:

39. The non-transitory computer readable medium of clause 37, wherein the method further comprises:

40. The non-transitory computer readable medium of clause 39, wherein the method further comprises:

41. The non-transitory computer readable medium of any one of clauses 37-40, wherein the inspection tool is an electron beam inspection tool.

The term "non-transitory medium" as used herein refers to any non-transitory medium that stores data and/or instructions that cause a machine to operate in a particular manner. Such non-transitory media may include non-volatile media and/or volatile media. Non-volatile media may include, for example, optical or magnetic disks, such as storage device 810. Volatile media may include dynamic memory, such as main memory 806. Non-transitory media include, for example, floppy disks, flexible disks, hard disks, solid state drives, magnetic tape or any other magnetic data storage medium, CD-ROMs, any other optical data storage medium, any physical medium with patterns of holes, RAMs, PROMs, and EPROMs, FLASH-EPROMs, NVRAMs, FLASH memory, registers, caches, any other memory chips or cartridges, and networked versions thereof.

Non-transitory media are different from, but may be used in conjunction with, transmission media. Transmission media may be involved in transferring information between storage media. For example, transmission media can include coaxial cables, copper wire and fiber optics, including the wires that comprise bus 802. Transmission media can also take the form of acoustic or light waves, such as those generated during radio wave and infrared data communications.

Various forms of media may be involved in carrying one or more sequences of one or more instructions to processor 804 for execution. For example, the instructions may initially be carried on a magnetic disk or solid state drive of a remote computer. The remote computer can load the instructions into its dynamic memory and send the instructions over a telephone line using a modem. A modem local to computer system 800 can receive the data on the telephone line and use an infra-red transmitter to convert the data to an infra-red signal. An infra-red detector can receive the data carried in the infra-red signal and appropriate circuitry can place the data on bus 802. Bus 802 carries the data to main memory 806, from which processor 804 retrieves and executes the instructions. The instructions received by main memory 806 may optionally be stored on storage device 810 either before or after execution by processor 804.

Computer system 800 may also include a communication interface 818 coupled to bus 802. Communication interface 818, such as communication interface 322 of defect classification server 320 or any communication interface (not shown) of knowledge recommendation server 330 of fig. 3, may provide bi-directional data communication coupled to a network link 820 that may be connected to a local network 822. For example, communication interface 818 may be an Integrated Services Digital Network (ISDN) card, a cable modem, a satellite modem, or a modem to provide a data communication connection to a corresponding type of telephone line. As another example, communication interface 818 may be a Local Area Network (LAN) card to provide a data communication connection to a compatible LAN. Wireless links may also be implemented. In any such implementation, communication interface 818 may send and receive electrical, electromagnetic or optical signals that carry digital data streams representing various types of information.

Network link 820 typically provides data communication through one or more networks to other data devices. For example, network link 820 may provide a connection through local network 822 to a host computer 824 or to data equipment operated by an Internet Service Provider (ISP) 826. ISP 826 in turn may provide data communication services through the world wide packet data communication network now commonly referred to as the "Internet" 828. Local network 822 and internet 828 both use electrical, electromagnetic or optical signals that carry digital data streams. The signals through the various networks and the signals on network link 820 and through communication interface 818, which carry the digital data to computer system 800 and from computer system 800, may be exemplary forms of transmission media.

Computer system 800 can send messages and receive data, including program code, through the network(s), network link 820 and communication interface 818. In the Internet example, a server 830 might transmit a requested code for an application program through Internet 828, ISP 826, local network 822 and communication interface 818.

The received code may be executed by processor 804 as it is received, and/or stored in storage device 810, or other non-volatile storage for later execution. In some embodiments, server 830 may provide information for display on a display.

According to the above disclosed embodiments, the defect inspection system includes a knowledge recommendation server, which may provide recommended knowledge files to a defect classification server, which may classify defects by using the knowledge files. In contrast to typical defect inspection systems, where a user needs to visually analyze a newly obtained wafer inspection image and spends a lot of time searching for knowledge files used in defect classification, the defect inspection system of the disclosed embodiments may perform the entire defect inspection process in a real-time scenario, i.e., immediately upon generation of the wafer inspection image, without user intervention. Thus, the throughput of the defect inspection process is increased.

In addition, typical defect inspection systems rely on the experience of the user to select knowledge files for use in defect classification, which may lead to inaccurate defect classification results. In contrast, the knowledge recommendation system of the disclosed embodiments searches for knowledge files based on defect feature parameters extracted from the defect patch images, which may generate more accurate results.

While the invention has been described in conjunction with various embodiments, other embodiments of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the invention being indicated by the following claims.

Claims

1. A server for knowledge recommendation for defect inspection, comprising:

a processor electronically coupled to an electronic storage device storing a plurality of knowledge files related to wafer defects, the processor configured to execute a set of instructions to cause the server to:

searching a knowledge file matched with the check image in the electronic storage device; and

transmitting the search results to the defect classification server causes the defect classification server to perform automatic defect classification or construct knowledge files, wherein each knowledge file includes a plurality of defect patch images and feature parameters for different types of defects in the same wafer process layer and generated by an inspection tool under the same inspection conditions.

2. The server of claim 1, wherein the processor is configured to execute the set of instructions to cause the server to:

acquiring a plurality of defect patch images from the knowledge file in the electronic storage device;

extracting feature parameters from the plurality of defect patch images and normalizing the feature parameters;

Generating a plurality of weighted representation models based on the normalized characteristic parameters, each weighted representation model representing a defect type; and

the weighted representation model is stored in the electronic storage device.

3. The server of claim 1, wherein the request for knowledge recommendation includes characteristic parameters of a plurality of defects identified in the inspection image, and

in searching for a knowledge file that matches the verification image, the processor is configured to execute the set of instructions to cause the server to:

searching a knowledge file having feature parameters matching the feature parameters of the plurality of defects identified in the newly obtained inspection image.

4. The server of claim 1, wherein the defect patch image is generated by an electron beam inspection tool.

5. A defect classification server for defect classification, comprising:

a processor configured to execute a set of instructions to cause the defect classification server to:

Receiving a test image of the wafer from a test tool;

receiving knowledge recommendation results from the knowledge recommendation server;

determining whether the knowledge recommendation result comprises a knowledge file; and

in response to determining that the knowledge recommendation includes a knowledge file, performing automatic defect classification on the inspection image by using the knowledge file,

wherein the defect classification processor is configured to execute a set of instructions to cause the server to:

6. The server of claim 5, wherein the processor is configured to execute a set of instructions to cause the server to:

after performing the automatic defect classification on the inspection image, prompting a user to inspect the results of the automatic defect classification, and/or

Wherein the processor is configured to execute a set of instructions to cause the server to:

7. The server of claim 5, wherein the inspection tool is an electron beam inspection tool.

8. A method for knowledge recommendation, comprising:

searching a knowledge file matched with the inspection image in an electronic storage device, wherein the electronic storage device stores a plurality of knowledge files related to wafer defects;

acquiring a plurality of defect patch images from the plurality of knowledge files; wherein the defect patch image is generated by an electron beam inspection tool; and

transmitting the search result to the defect classification server so that the defect classification server performs automatic defect classification or constructs a knowledge file.

9. The method of claim 8, further comprising:

the weighted representation model is stored in the electronic storage device.

10. The method of claim 8, wherein the request for knowledge recommendation includes characteristic parameters of a plurality of defects identified in the inspection image, and

searching the knowledge file further includes:

11. The method of claim 8, wherein each knowledge file includes a plurality of defect patch images and feature parameters for different types of defects in the same wafer process layer and generated by an inspection tool under the same inspection conditions.